4.1 Experimental Procedures
In this thesis, we employed the structure of the superstrate p-i-n hydrogenated microcrystalline silicon solar cells, and the solar cell structure employed was glass (Asahi's textured TCO Glass)/front ZnO/p-i-n/back ZnO/Ag with an active of 0.25 cm2. The device deposition condition: front ZnO contact layer was deposited ZnO:Ga with 40nm thickness by sputter. All (μc-Si:H) p, i, and n layers were subsequently deposited by 40 MHz in the very high frequency plasma enhanced chemical vapor deposition (VHF-PECVD) multi-chamber system. The window layer (P-layer) was deposited μc-Si:B with 15 nm thickness. The intrinsic layer was deposited μc-Si with 1500 nm thickness. The n-layer was deposited μc-Si:P with 15 nm thickness. Back ZnO contact layer was deposited ZnO:Ga with 100 nm thickness by sputter. Back Silver metal layer was deposited Ag with 200 nm thickness. In VHF-PECVD process, we use a H2 and SiH4 gases. Application of high-rate μc-Si:H deposition was most feasible to promote electron-impact dissociation of SiH4 and H2 into SiH2 (SiH H SiH ). The dilution ratios R H /SiH were 1000/22 and 1000/14 and 1000/13 for μc-Si:H intrinsic layer at pressures of 4Torr、6Torr、8Torr, respectively.
The solar cells were characterized by current–voltage (J–V) and spectral response measurements [40] under standard air mass 1.5 (100 mW/cm2) illuminations.
The current-voltage characteristic measurement of thin film solar cell devices was performed by Agilent 4156, capacitance profiling by drive level capacitance profiling methods have been carried out by using an HP 4284A impedance meter. Otherwise, we make the conductivity and activation energy measurement of μc-Si:H intrinsic layer, and measurement flow：the voltage sweep 0V to 10V and select a constant
20
voltage, V (mostly 10 V) is applied over the contacts and the current, I, is measured.
We use equation of dark conductivity,
σ I W
V L D where I is measured current
V is applied voltage W is width of the film L is length of the film D is thickness of the film
to find dark conductivity. We make the sample to heat from 300K to 370K for temperature measurements, and employ equation of activation energy,
σ σ Exp E
KT where σ is the coefficient of conductivity
Ea is activation energy to find activation energy of the sample.
4.2 Results and Discussion
Figure 4.1 shows current-voltage characteristics of microcrystalline silicon thin film solar cells under standard illumination conditions. These devices were deposited in different pressures. We could observe that short circuit current density (J ) and conversion efficiency (η) increased when deposition pressure increased, and shunt resistance (R ) and series resistance (R ) decrease when deposition pressure increase, but open circuit voltage (V ) was almost same. Table 4-1 summarizes the illuminated J–V parameters of μc-Si:H p–i–n solar cells obtained for varied i layer deposition rates.
Figure 4.2 shows the Quantum efficiency (QE) for several high-efficiency thin
21
film solar cells. Spectral response measurements are valuable to characterize the photocurrent and are commonly used to determine the losses responsible for reducing the measured short circuit current density from the maximum achievable photocurrent.
QE is a dimensionless parameter given by the number of electrons which exit the device per incident photon at each wavelength. Device losses measured by QE can be optical, due to the front reflection and absorption in the window, transparent conductor, and other layers, or electronic, due to recombination losses in the absorber.
We could observe that at pressures of 4、6Torr and pressures of 8Torr which corresponded to QE were approximately the same, respectively, in short wavelength region. Because we varied thicknesses of deposited p-type microcrystalline silicon layer, and the deposition thicknesses were thinner slightly at pressures of 8Torr than at pressures of 4、6Torr. They could decrease absorption losses in the thinner thickness, and could increase the number of electrons which exit the device per incident photon. Hence, the thinner thicknesses were higher QE. On the other hand, in long wavelength region, we could observe that the deposition pressure increased and QE also increased. The reason that lower deposition pressure of the thin films would deposited rate slowly, result in more oxygen go into the thin films, and oxygen could generate electron as donors when it went into the thin films, then oxygen would be trapped in the vicinity of grain boundaries.[41] Hence, the thin films resulted in tending to n- quality gradually. The result that almost made built in field totally locate in the vicinity of p/i interface, and caused built in field to become weak in the vicinity of i-layer bulk and i/n interface, as shown in Figure 4.3. Because the photocurrent was proportional to built in field, the thin films of lower deposition pressure were lower photocurrent.
In order to verify high-rate deposition can decrease oxygen go into the film, we make the conductivity and activation energy measurement of μc-Si intrinsic layer.
22
Figure 4.4 shows dark conductivity and photo conductivity as a function of deposition pressure. We observe that with dark conductivity and photo conductivity decrease as deposition pressure increase. It indicates obviously smaller resistance and better conductivity for low deposition pressure, and characteristic of the film is tending to n -quality. Figure 4.5 shows activation energy as a function of deposition pressure. We observe that with activation energy increases as deposition pressure increases, and the value of activation energy is close to 550 meV for 8Torr deposition pressure, as shown in Table4-2. Energy band gap of μc-Si film is about 1.12eV and midgap of μc-Si film is about 550 meV, and activation energy is close to 550 meV for 8Torr deposition pressure, so Fermi level EF of μc-Si film is close to Ei for high deposition pressure. Yet, the value activation energy is 478 meV and Fermi level EF of μc-Si film is above Ei for 4Torr deposition pressure. Therefore, characteristic of μc-Si film is tending to n-quality for 4Torr deposition pressure. Hence, high deposition pressure not only made characteristic of the film tending to intrinsic quality but also could decrease oxygen go into the film.
In brief, the thinner thicknesses of deposited p-type microcrystalline silicon layer were higher QE and independent on deposition pressures in short wavelength region;
lower deposition pressure of the thin films would more oxygen go into the thin films, then the thin films made tending to n- quality gradually result in almost make electric field totally locate in the vicinity of p/i interface, and the photocurrent was proportional to built in field. Therefore, the thin films of lower deposition pressure were lower photocurrent in long wavelength region.
In addition, in order to investigate about the relationship between the spatial distribution of defects in the intrinsic layer and the deposition pressure, we would make use of a technique of drive-level capacitance profiling (DLCP) for varied deposition pressures.
23
Figure 4.6(a)-4.8(a) show typical capacitance-voltage curve of the variation of junction capacitance with 100Hz for deposition pressures of 4-8Torr at room temperature, respectively. Figure 4.6(b)-4.8(b) show NDL versus depletion width for deposition pressures of 4-8Torr at room temperature, respectively. We could observe that number of defects in the vicinity of the intrinsic layer of p/i interface were lower for higher deposition pressures; however, number of defects in the intrinsic layer of bulk and i/n interface were not any variation explicitly for arbitrarily deposition pressures, as shown in Figure 4.9. Because gas of the lower deposition pressures was less reactants than the higher deposition pressures and generated less precursors, and there had prone to plasma ion bombardment more probability in the vicinity of p/i interface, result in the destruction of the interface is very serious. Hence, it is better in the interface for higher pressure deposition. [42]
Figure 4.10 show the measured values of current-voltage characteristics for μc-Si thin film solar cells with temperature measurement from 25℃ to 85℃ under the condition of darkness. We could observe that current density increase as the temperature of solar cells increase for any temperature condition. Under the forward bias, we would observe change in current density rise are different in the voltage region of 0.2V~0.4V and 0.6~1V. The characteristics of solar cells and pn diodes are the same under the condition of darkness. The current-voltage characteristics of ideal diodes are given by
J J Exp KTV 1 (1) where Jo is the saturation current density
n is the ideality factor.
When the ideal diffusion current dominates, n equals 1; whereas when the recombination current dominates, n equals 2.Therefore, we attempt to extract the
24
value of n by fitting from eq. (1) in the voltage region of 0.2V~0.4V and0.6~1V, respectively, and shown in Table 4-3. Jo increased with increasing temperature for all cells, and the value of n is 2 approximately for all cells in the voltage region of 0.2V~0.4V. The result show temperature dependence of Jo and ideality factor n equals 2 approximately represents the quality of the recombination center. So that is recombination current dominant in the voltage region of 0.2V~0.4V. In the voltage region of 0.6~1V, Jo also increased with increasing temperature for all cells, so temperature dependence of Jo. However, the value of n is greater than 1 for all cells.
This phenomenon is associated with series resistance effect. At low current level, the IR drop across the neutral regions is usually small compared to kT/q (26 mV at 300K),
where I is the forward current and R is the series resistance. At high current level, the IR drop across the neutral regions is very large, and the neutral regions of i layer is
huge specifically. Hence, current increases more gradually with forward voltage.
Figure 4.11 show the measured values of current-voltage characteristics for μc-Si thin film solar cells with temperature measurement from 25℃ to 85℃ under standard illumination conditions. We could observe with the value of Voc decrease as the temperature increase, as shown in Figure 4.12; with the value of Jsc increase as the temperature increase, as shown in Figure 4.13; with the value of conversion efficiency decrease as the temperature increase, as shown in Figure 4.14. The first, Voc is associated with energy band gap Eg, is given by
eV EF EF E kTln G
bRN N
Eg is associated with temperature, and when temperature is rise with Eg is dropped.
Therefore, with the value of Voc decrease when temperature is rise. Device of 4Torr deposition pressure reduce the maximum amount of Voc with temperature increase to 85oC and the value about 11.5%, as shown in table 4-4. The second, Jsc is associated
25
with collect electron-hole pairs capability. Because they generate a mass of electron-hole pairs in the intrinsic layer as temperature is rise, it can collect a mass of electron-hole pairs output abundant currents. Therefore, with the value of Jsc increase when temperature is rise. Device of 8Torr deposition pressure increase the maximum amount of Jsc with temperature increase to 85oC and the value about 2.94%. The last, conversion efficiency is associated with Voc and Jsc. Although with the value of Jsc increase as temperature is rise, with the value of Voc decrease more than the value of Jsc increase as temperature is rise. The whole of conversion efficiency decayed as temperature rise. Device of 4Torr deposition pressure decrease the maximum amount of conversion efficiency with temperature increase to 85oC and the value about 14.8%.
According to the above results, Device of a-SiO of p layer is the most sensitive to temperature.
26